- Add setup instructions to the README (#29)

README.md

@@ -15,6 +15,22 @@ Sentiment analysis is performed on the comment text using the 'nltk' (Natural
 Language Toolkit) python library. This data is added to the database entries of
 each comment.
 
+## Setup Instructions
+1. Install `python3`
+2. Create your python virtual environment with `python3 -m venv <env name>`
+3. Install the packages in the `src/requirements.txt` file to your virtual
+   environment with `pip install -r requirements.txt`.
+4. Create a file in the `/src` directory called `.env`. This file should contain
+   The following values:
+   ```
+    API_KEY=<key> # your YouTube Data API key
+    DB_FILE=<filename> # the database file to which collected data will be written
+    LOG_LEVEL=[debug|info|warn|error]
+    ```
+    For information about how to create an API key, see [here](https://blog.hubspot.com/website/how-to-get-youtube-api-key).
+5. Run the tool with `python astro.py <YouTube video URL>` to start collecting
+   data. You can see output from an example run in the next section.
+
 ## Example
 This output below was generated by providing the video ID string of a relatively
 small YouTube channel.
@@ -34,31 +50,14 @@ DEBUG:astro.py:main: Collected data preview:
 
 ## Background
 YouTube has been a primary source of information and entertainment in my house
-for years now. Recently when looking at the comments of various videos, I've
-noticed what I feel is strange behavior. However, I am not someone who regularly
-engages with social media in that way, so this project is my attempt to
-investigate and understand the behavior I see online.
+for years. I've found that when reading comments on YouTube videos, I'm often 
+perplexed by the content there. Wanting to understand this behavior, whether it
+was the product of real users or bots, I started researching social media usage.
+This project is my attempt to gather data from YouTube videos and their comments
+in order to analyze trends in the data, if any, in an effort to better
+understand YouTube commenting behavior and its impact on video performance.
 
 The name 'Astro' was chosen as a short form of 'Astroturf', a term used to
 describe artificial social movements, since I was initially working toward
 identifying bot campaigns. I've since decided to restrict the scope of the
 project (at least for now), since that goal will require much more research.
-
-## Prerequisites
-These packages are required in order to run this tool.
-
-- YouTube Data API
-Used for accessing YouTube data. Install with:
-`pip install google-api-python-client`
-
-- dotenv
-Used for loading environment variables from a local .env file. Install with:
-`pip insatll python-dotenv`
-
-- pandas
-Used for capturing YouTube data in dataframes (for now). Install with:
-`pip install pandas`
-
-- Natural Language Toolkit
-Used for sentiment analysis of comments. Install with:
-`pip insatll nltk`